Pypethon: Design of an Multibody Pipe Inspection Robot Using Artificial Vision

Rivadeneira F.; Cuellar F. · 2025 11th International Conference on Mechatronics and Robotics Engineering, ICMRE 2025 (2025) · DOI: 10.1109/icmre64970.2025.10976295

Resumen

Natural gas plays a crucial role in daily life, powering many appliances and equipment. Its importance is evident in rising energy consumption, as noted by regulatory and global energy agencies, highlighting the continued growth in both local and worldwide demand. Due to this, natural gas pipelines require regular and effective maintenance to ensure quality distribution and prevent potential accidents like leaks. To address this issue, this study proposes the design of an 8-inch pipeline inspection robot capable of detecting faults on the internal surfaces of the pipes, such as rust and corrosion. As a result, the robot, along with the localization and defect detection algorithms, was implemented, achieving a Mean Absolute Error of 0.5 for the localization algorithm and F1-Scores of 67.26% and 69.23% for the rust and corrosion detection algorithms, respectively, using the YOLOv9 model.

Instituciones: Pontificia Universidad Católica del Perú